The detection of specific nucleic acid sequences is an important tool for diagnostic medicine and molecular biology research. Gene probe assays currently play roles in identifying infectious organisms such as bacteria and viruses, in probing the expression of normal genes and identifying mutant genes such as oncogenes, in typing tissue for compatibility preceding tissue transplantation, in matching tissue or blood samples for forensic medicine, and for exploring homology among genes from different species.
Ideally, a gene probe assay should be sensitive, specific and easily automatable (for a review, see Nickerson, Current Opinion in Biotechnology 4:48-51 (1993)). The requirement for sensitivity (i.e. low detection limits) has been greatly alleviated by the development of the polymerase chain reaction (PCR) and other amplification technologies which allow researchers to amplify exponentially a specific nucleic acid sequence before analysis (for a review, see Abramson et al., Current Opinion in Biotechnology, 4:41-47 (1993)).
Specificity, in contrast, remains a problem in many currently available gene probe assays. The extent of molecular complementarity between probe and target defines the specificity of the interaction. Variations in the concentrations of probes, of targets and of salts in the hybridization medium, in the reaction temperature, and in the length of the probe may alter or influence the specificity of the probe/target interaction.
It may be possible under some limited circumstances to distinguish targets with perfect complementarity from targets with mismatches, although this is generally very difficult using traditional technology, since small variations in the reaction conditions will alter the hybridization. New experimental techniques for mismatch detection with standard probes include DNA ligation assays where single point mismatches prevent ligation and probe digestion assays in which mismatches create sites for probe cleavage.
Finally, the automation of gene probe assays remains an area in which current technologies are lacking. Such assays generally rely on the hybridization of a labeled probe to a target sequence followed by the separation of the unhybridized free probe. This separation is generally achieved by gel electrophoresis or solid phase capture and washing of the target DNA, and is generally quite difficult to automate easily.
The time consuming nature of these separation steps has led to two distinct avenues of development. One involves the development of high-speed, high-throughput automatable electrophoretic and other separation techniques. The other involves the development of non-separation homogeneous gene probe assays.
For example, Gen-Probe Inc., (San Diego, Calif.) has developed a homogeneous protection assay in which hybridized probes are protected from base hydrolysis, and thus are capable of subsequent chemiluminescence. (Okwumabua et al. Res. Microbiol. 143:183 (1992)). Unfortunately, the reliance of this approach on a chemiluminescent substrate known for high background photon emission suggests this assay will not have high specificity. EPO application number 86116652.8 describes an attempt to use non-radiative energy transfer from a donor probe to an acceptor probe as a homogeneous detection scheme. However, the fluorescence energy transfer is greatly influenced by both probe topology and topography, and the DNA target itself is capable of significant energy quenching, resulting in considerable variability. Therefore there is a need for DNA probes which are specific, capable of detecting target mismatches, and capable of being incorporated into an automated system for sequence identification.
As outlined above, molecular biology relies quite heavily on modified or labeled oligonucleotides for traditional gene probe assays (Oligonucleotide Synthesis: A Practical Approach. Gait et al., Ed., IRL Press: Oxford, UK, 1984; Oligonucleotides and Analogues: A Practical Approach. Ed. F. Eckstein, Oxford University Press, 1991). As a result, several techniques currently exist for the synthesis of tailored nucleic acid molecules. Since nucleic acids do not naturally contain functional groups to which molecules of interest may easily be attached covalently, methods have been developed which allow chemical modification at either of the terminal phosphates or at the heterocyclic bases (Dreyer et al. Proc. Natl. Acad. Sci. USA, 1985, 82:968).
For example, analogues of the common deoxyribo- and ribonucleosides which contain amino groups at the 2' or 3' position of the sugar can be made using established chemical techniques. (See Imazawa et al., J. Org. Chem., 1979, 44:2039; Imazawa et al., J. Org. Chem. 43(15):3044 (1978); Verheyden et al., J. Org. Chem. 36(2):250 (1971); Hobbs et al., J. Org. Chem. 42(4):714 (1977)). In addition, oligonucleotides may be synthesized with 2'-5' or 3'-5' phosphoamide linkages (Beaucage et al., Tetrahedron 49(10):1925 (1992); Letsinger, J. Org. Chem., 35:3800 (1970); Sawai, Chem. Lett. 805 (1984); Oligonucleotides and Analogues: A Practical Approach, F. Eckstein, Ed. Oxford University Press (1991)).
The modification of nucleic acids has been done for two general reasons: to create nonradioactive DNA markers to serve as probes, and to use chemically modified DNA to obtain site-specific cleavage.
To this end, DNA may be labeled to serve as a probe by altering a nucleotide which then serves as a replacement analogue in the nick translational resynthesis of double stranded DNA. The chemically altered nucleotide may then provide reactive sites for the attachment of immunological or other labels such as biotin. (Gilliam et al., Anal. Biochem. 157:199 (1986)). Another example uses ruthenium derivatives which intercalate into DNA to produce photoluminescence under defined conditions. (Friedman et al., J. Am. Chem. Soc. 112:4960 (1990)).
In the second category, there are a number of examples of compounds covalently linked to DNA which subsequently cause DNA chain cleavage. For example 1,10-phenanthroline has been coupled to single-stranded oligothymidylate via a linker which results in the cleavage of poly-dA oligonucleotides in the presence of Cu.sup.2+ and 3-mercaptopropionic acid (Francois et al., Biochemistry 27:2272 (1988)). Similar experiments have been done for EDTA.sup.1 -Fe(II) (both for double stranded DNA (Boutorin et al., FEBS Lett. 172:43-46 (1986)) and triplex DNA (Strobel et al., Science 249:73 (1990)), porphyrin-Fe(III) (Le Doan et al., Biochemistry 25:6736-6739 (1986)), and 1,10-phenanthronine-Cu(I) (Chen et al., Proc. Natl. Acad. Sci USA, 83:7147 (1985)), which all result in DNA chain cleavage in the presence of a reducing agent in aerated solutions. A similar example using porphyrins resulted in DNA strand cleavage, and base oxidation or cross-linking of the DNA under very specific conditions (Le Doan et al., Nucleic Acids Res. 15:8643 (1987)).
Other work has focused on chemical modification of heterocyclic bases. For example, the attachment of an inorganic coordination complex, Fe-EDTA, to a modified internal base resulted in cleavage of the DNA after hybridization in the presence of dioxygen (Dreyer et al., Proc. Natl. Acad. Sci. USA 82:968 (1985)). A ruthenium compound has been coupled successfully to an internal base in a DNA octomer, with retention of both the DNA hybridization capabilities as well as the spectroscopic properties of the ruthenium label (Telser et al., J. Am. Chem. Soc. 111:7221 (1989)). Other experiments have successfully added two separate spectroscopic labels to a single double-stranded DNA molecule (Telser et al., J. Am. Chem. Soc. 111:7226 (1989)).
The study of electron transfer reactions in proteins and DNA has also been explored in pursuit of systems which are capable of long distance electron transfer.
To this end, intramolecular electron transfer in protein-protein complexes, such as those found in photosynthetic proteins and proteins in the respiration pathway, has been shown to take place over appreciable distances in protein interiors at biologically significant rates (see Bowler et al., Progress in Inorganic Chemistry: Bioinorganic Chemistry, Vol. 38, Ed. Stephen J. Lippard (1990). In addition, the selective modification of metalloenzymes with transition metals has been accomplished and techniques to monitor electron transfer in these systems developed. For example, electron transfer proteins such as cytochrome c have been modified with ruthenium through attachment at several histidines and the rate of electron transfer from the heme Fe.sup.2+ to the bound Ru.sup.3+ measured. The results suggest that electron transfer "tunnel" pathways may exist. (Baum, Chemical & Engineering News, Feb. 22, 1993, pages 2023; see also Chang et al., J. Am. Chem. Soc. 113:7056 (1991)). In related work, the normal protein insulation, which protects the redox centers of an enzyme or protein from nondiscriminatory reactions with the exterior solvent, was "wired" to transform these systems from electrical insulators into electrical conductors (Heller, Acc. Chem. Res. 23:128 (1990)).
There are a few reports of photoinduced electron transfer in a DNA matrix. In these systems, the electron donors and acceptors are not covalently attached to the DNA, but randomly associated with the DNA, thus rendering the explicit elucidation and control of the donor-acceptor system difficult. For example, the intense fluorescence of certain quaternary diazoaromatic salts is quenched upon intercalation into DNA or upon exposure to individual mononucleotides, thus exhibiting electron donor processes within the DNA itself. (Brun et al., J. Am. Chem. Soc. 113:8153 (1991)).
Another example of the difficulty of determining the electron transfer mechanism is found in work done with some photoexcitable ruthenium compounds. Early work suggested that certain ruthenium compounds either randomly intercalate into the nucleotide bases, or bind to the helix surface. (Purugganan et al., Science 241:1645 (1988)). A recent reference indicates that certain ruthenium compounds do not intercalate into the DNA (Satyanarayana et al., Biochemistry 31(39):9319 (1992)); rather, they bind non-covalently to the surface of the DNA helix.
In these early experiments, various electron acceptor compounds, such as cobalt, chromium or rhodium compounds were added to certain DNA-associated ruthenium electron donor compounds. (Puragganan et al., Science 241:1645 (1988); Orellana et al., Photochem. Photobiol. 499:54 (1991); Brun et al., J. Am. Chem. Soc. 113:8153 (1991); Davis, Chem-Biol. Interactions 62:45 (1987); Tomalia et al., Acc. Chem. Res., 24:332 (1991)). Upon addition of these various electron acceptor compounds, which randomly bind non-covalently to the helix, quenching of the photoexcited state through electron transfer was detected. The rate of quenching was dependent on both the individual electron donor and acceptor as well as their concentrations, thus revealing the process as bimolecular.
In one set of experiments, the authors postulate that the more mobile surface bound donor promotes electron transfer with greater efficiency than the intercalated species, and suggest that the sugar-phosphate backbone of DNA, and possibly the solvent medium surrounding the DNA, play a significant role in the electron transport. (Purugganan et al., Science 241:1645 (1988)). In other work, the authors stress the dependence of the rate on the mobility of the donor and acceptor and their local concentrations, and assign the role of the DNA to be primarily to facilitate an increase in local concentration of the donor and acceptor species on the helix. (Orellana et al., supra).
In another experiment, an electron donor was reportedly randomly intercalated into the stack of bases of DNA, while the acceptor was randomly associated with the surface of the DNA The rate of electron transfer quenching indicated a close contact of the donor and the acceptor, and the system also exhibits enhancement of the rate of electron transfer with the addition of salt to the medium. (Fromherz et al., J. Am. Chem. Soc. 108:5361 (1986)).
In all of these experiments, the rate of electron transfer for non-covalently bound donors and acceptors is several orders of magnitude less than is seen in free solution.
An important stimulus for the development of long distance electron transfer systems is the creation of synthetic light harvesting systems. Work to date suggests that an artificial light harvesting system contains an energy transfer complex, an energy migration complex, an electron transfer complex and an electron migration complex (for a topical review of this area, see Chemical & Engineering News, Mar. 15, 1993, pages 38-48). Two types of molecules have been tried: a) long organic molecules, such as hydrocarbons with covalently attached electron transfer species, or DNA, with intercalated, partially intercalated or helix associated electron transfer species, and b) synthetic polymers.
The long organic molecules, while quite rigid, are influenced by a number of factors, which makes development difficult These factors include the polarity and composition of the solvent, the orientation of the donor and acceptor groups, and the chemical character of either the covalent linkage or the association of the electron transfer species to the molecule.
The creation of acceptable polymer electron transfer systems has been difficult because the available polymers are too flexible, such that several modes of transfer occur. Polymers that are sufficiently rigid often significantly interfere with the electron transfer mechanism or are quite difficult to synthesize.
Thus the development of an electron transfer system which is sufficiently rigid, has covalently attached electron transfer species at defined intervals, is easy to synthesize and does not appreciably interfere with the electron transfer mechanism would be useful in the development of artificial light harvesting systems.
In conclusion, the random distribution and mobility of the electron donor and acceptor pairs, coupled with potential short distances between the donor and acceptor, the loose and presumably reversible association of the donors and acceptors, the reported dependence on solvent and broad putative electron pathways, and the disruption of the DNA structure of intercalated compounds rendering normal base pairing impossible all serve as pronounced limitations of long range electron transfer in a DNA matrix. Therefore, a method for the production of rigid, covalent attachment of electron donors and acceptors to provide minimal perturbations of the nucleic acid structure and retention of its ability to base pair normally, is desirable. The present invention serves to provide such a system, which allows the development of novel bioconductors and diagnostic probes.